No user-submitted related servers found.
Scored across 19 tools
Most tools target a distinct resource+action, and descriptions actively police the boundaries (video_find_phrase for exact wording vs video_search for meaning; video_ask for evidence vs video_search for ranked hits). The weaker spots are video_status / video_info / video_probe, which all read as 'tell me about something' and require reading descriptions to separate system readiness, indexed-video metadata and pre-download link inspection.
Every tool carries the same `video_` prefix, so the namespace is instantly recognizable and scannable. However, the suffix alternates between verbs/actions (download, ingest, ask, search, delete, cancel) and nouns (status, info, transcript, frames, chapters), and video_job vs video_jobs differs only by plurality, which is a minor readability wart rather than a real inconsistency.
19 tools is on the heavier side of the ideal band, but the domain genuinely spans an ingest pipeline, job management, retrieval (three search channels), transcript/frame/chapter inspection, and clipping, so nearly every tool earns its slot. The one arguable redundancy is video_download sitting alongside video_ingest, though they serve different ends (get a file vs build a searchable index).
The lifecycle is well covered: probe → download/ingest/batch → job status/list/cancel → list/info → search/ask/transcript/frames/chapters/clip → delete. Remaining gaps are minor: there is no corpus-level or cross-video question answering despite batch ingest, no re-index/refresh of a single video, and no way to prune job history, but agents can work around all of these.